Improved Modelling of Electric Loads for Enabling Demand

Improved Modelling of Electric Loads for
Enabling Demand Response by Applying
Physical and Data-Driven Models - RESPONSE
The aim of the project is to develop enhanced models for load and control response forecasting
required by dynamic on-line optimization of demand response (DR) actions and network operation
in a future sustainable energy system. DR is one of the key issues in adaptation of energy
production and consumption and in creating flexibility to integrated energy systems. In addition to
DR also other novel forthcoming Smart Grid related functions like price elasticity, customer level
micro generation and electric vehicles affects distribution network loads and these effects should be
taken into account in the load models.
The project is carried out in the multi-disciplinary research consortium with wide scientific
expertise including units from Tampere University of Technology, University of Eastern Finland
and VTT.
Research project focuses on the following objectives:
1) Developing improved models for short term load forecasting and optimization of electric load
dynamic responses to load control actions and weather variations.
2) Analyzing and developing criteria for comparing the performance of the short term load
forecasting.
3) Enhancing utilization of smart metering data and other spatial information in updating and
verification of load response models.
4) Proving that the developed load modelling methods improve the state estimation and network
analysis accuracy.
The research methods include various mathematical and statistical methods, including artificial
neural networks, support vector machines, Bayesian methods, Kalman filtering, modern control and
optimization methods, and simulation environments.
The research project utilizes industrial co-operation in terms of data and knowledge exchange and
practical piloting. Large smart metering data sets and field tests from several network companies
are used together with open data (i.e. weather, building and socioeconomic grid data) in modelling.
Building, weather and
socioeconomic data (derived
from opening public registers)
Measurements from smart
meters and/or substations,
and field tests
Enhanced utilisation of smart metering
data and other datasets (Objective 3)
Development of improved load and response forecast
models based on data-driven and physical modeling
methodologies (Objective 1)
-can forecast accurately in different situations
-enhanced capability to adapt to changes in load behavior
-are easy to maintain and update
-
Evaluation and criteria for
the performance of load
forecasting (Objective 2)
Forecasted load and responses in different meteorological conditions,
physical and socioeconomic environments and control scenarios
Research group (titles and tasks, and contact information)
Tampere University of Technology (TUT):
 Professor Pertti Järventausta, (leader of the whole project), pertti.jarventausta@tut.fi
 Professor Hannu Koivisto (modern control methods and multiobjective optimization),
hannu.koivisto@tut.fi
 M.Sc. Antti Mutanen (load modelling and state estimation), antti.mutanen@tut.fi
 M.Sc. Antti Rautiainen (modelling of demand response and effects of electric vehicles),
antti.rautiainen@tut.fi
 M.Sc. Tao Chen (methods for indicating changes in loads), tao.chen@tut.fi
University of Eastern Finland (UEF):
 Professor Mikko Kolehmainen (leader of UEF’s part), mikko.kolehmainen@uef.fi
 Ph.D. Harri Niska (neural network based load models and feature selection), harri.niska@uef.fi
 Ph.D. Mauno Rönkkö (external consulting in data semantics and fusion), mauno.ronkko@uef.fi
 M.Sc., Ph.D. student Jukka Saarenpää (spatial load modelling), jukka.saarenpaa@uef.fi
VTT:
 Dr. Tech Seppo Hänninen (leader of VTT’s part), seppo.hanninen@vtt.fi
 Dr. Tech. Pekka Koponen, (modelling, forecasting and optimization of short term control
responses of loads and DER), pekka.koponen@vtt.fi
 M.Sc. Göran Koreneff (modelling of loads and DR and their value in energy markets),
goran.koreneff@vtt.fi